Machine Learning

1 readers

1 users here now

Community Rules:

Be nice. No offensive behavior, insults or attacks: we encourage a diverse community in which members feel safe and have a voice.
Make your post clear and comprehensive: posts that lack insight or effort will be removed. (ex: questions which are easily googled)
Beginner or career related questions go elsewhere. This community is focused in discussion of research and new projects that advance the state-of-the-art.
Limit self-promotion. Comments and posts should be first and foremost about topics of interest to ML observers and practitioners. Limited self-promotion is tolerated, but the sub is not here as merely a source for free advertisement. Such posts will be removed at the discretion of the mods.

founded 2 years ago

MODERATORS

communick@academy.garden

[P] Yolov8 hyperparameters for best generalization (alien.top)

submitted 2 years ago by Banntu@alien.top to c/machinelearning@academy.garden

5 comments fedilink hide all child comments

I’m training an object detection model on yolov8 but my training data is a little biased because it doesn’t represent the real life distribution.
(I want to count objects of one class but different shape in a video and need them to be detected with near equal probability. ) How can I make sure to generalise the model enough so that the bias doesn’t have too much of an effect? I know it will come with more false positives, but that’s not a problem.

you are viewing a single comment's thread
view the rest of the comments

[–] Banntu@alien.top 1 points 2 years ago (1 children)

Yeah I thought that might be the case.
The projects goal is the following:

3D-Camera capturing live images of fish
yolov8 pose model detecting fish that are completely visible and not facing the camera, also detecting nose and tail of the fish
using the distance values of 3D-camera to calculate length of fish in mm
running the model on a live video stream to estimate the distribution of fish-lengths in a fish-tank

The problem is the following:
I have annotated images of multiple growth stages of fish, but the average growth stage of the fish in the training data will almost always be either smaller or bigger than the ones im measuring.
So when I'm training a model on all data I have and then running the model in a tank of fish that are at the upper end of growth, than the model will detect the smaller fish inside that tank more often, because most fish in the training data are smaller then the fish in the tank.
Does that make sense?

These values are just to show what I mean (Expecting that the model is always trained on all 5k samples)

growth stage	1 month	2 months	3 months	4 months	5 months
training samples	1000	1000	1000	1000	1000
length error on real live data	+ 10%	+ 5 %	0 %	- 5%	-10%

[–] Toilet2000@alien.top 1 points 2 years ago

If that’s not something that you already do, use data augmentation. Especially scaling.

It might not help for features related to age (such as color changes and the likes), but it can definitely help remove the relative size bias, especially if your dataset was created in a single scenario with a fixed camera distance.

If you do know some features of "an older fish", you can also apply those transformations on masks of younger fish from your first trained but biaised model. Somewhat like a "semi-synthetic" data augmentation.

For example, if older fishes are browner, you can skew the hue of the masks by a certain amount to get brownish young fishes.